Aromatic thiol-mediated cleavage of N-O bonds enables chemical ubiquitylation of folded proteins

Abstract

Access to protein substrates homogenously modified by ubiquitin (Ub) is critical for biophysical and biochemical investigations aimed at deconvoluting the myriad biological roles for Ub. Current chemical strategies for protein ubiquitylation, however, employ temporary ligation auxiliaries that are removed under harsh denaturing conditions and have limited applicability. We report an unprecedented aromatic thiol-mediated N-O bond cleavage and its application towards native chemical ubiquitylation with the ligation auxiliary 2-aminooxyethanethiol. Our interrogation of the reaction mechanism suggests a disulfide radical anion as the active species capable of cleaving the N-O bond. The successful semisynthesis of full-length histone H2B modified by the small ubiquitin-like modifier-3 (SUMO-3) protein further demonstrates the generalizability and compatibility of our strategy with folded proteins.

ARTICLE

Received 26 Aug 2015 | Accepted 22 Aug 2016 | Published 29 Sep 2016

Caroline E. Weller1, Abhinav Dhall1, Feizhi Ding1, Edlaine Linares2, Samuel D. Whedon1, Nicholas A. Senger1, Elizabeth L. Tyson1, John D. Bagert3, Xiaosong Li1, Ohara Augusto2 & Champak Chatterjee1

Access to protein substrates homogenously modied by ubiquitin (Ub) is critical for biophysical and biochemical investigations aimed at deconvoluting the myriad biological roles for Ub. Current chemical strategies for protein ubiquitylation, however, employ temporary ligation auxiliaries that are removed under harsh denaturing conditions and have limited applicability. We report an unprecedented aromatic thiol-mediated NO bond cleavage and its application towards native chemical ubiquitylation with the ligation auxiliary 2-aminooxyethanethiol. Our interrogation of the reaction mechanism suggests a disulde radical anion as the active species capable of cleaving the NO bond. The successful semisynthesis of full-length histone H2B modied by the small ubiquitin-like modier-3 (SUMO-3) protein further demonstrates the generalizability and compatibility of our strategy with folded proteins.

DOI: 10.1038/ncomms12979 OPEN

Aromatic thiol-mediated cleavage of NO bonds enables chemical ubiquitylation of folded proteins

1 Department of Chemistry, University of Washington, Seattle, Washington 98195, USA. 2 Departamento de Bioqumica, Instituto de Qumica-Universidade de Sao Paulo, Sao Paulo 05513-970, Brazil. 3 Department of Chemistry, Princeton University, Princeton, New Jersey 08544, USA. Correspondence and requests for materials should be addressed to X.L. (email: mailto:[email protected]

Web End [email protected] ) or to O.A. (email:mailto:[email protected]

Web End [email protected] r) or to C.C. (email:mailto:[email protected]

Web End [email protected] u).

NATURE COMMUNICATIONS | 7:12979 | DOI: 10.1038/ncomms12979 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12979

The reversible conjugation of proteins with ubiquitin and ubiquitin-like proteins is a post-translational modication conserved in all eukaryotic organisms1. The ubiquitin

family consists of about 25 proteins, a majority of which can be conjugated with protein targets, either at specic Lys side-chain amines or at the protein N terminus2. For ubiquitin alone, there exist over 600 ligases that attach it to protein substrates in humans3. The many roles for protein modication by ubiquitin, termed ubiquitylation, include protein degradation, organelle-specic localization and the regulation of protein function4. Ubiquitylation is a dynamic modication and only a small fraction of proteins are ubiquitylated at any given time5, which complicates their isolation and hinders subsequent biochemical or biophysical studies aimed at unravelling the specic roles for ubiquitin.

In the last decade, a handful of chemical strategies have emerged that enable access to site-specically ubiquitylated proteins or close analogs thereof6,7. These approaches, however, rely on challenging multi-step synthetic strategies, harsh denaturing conditions and/or a desulfurization of the nal ubiquitylated product that is incompatible with cysteine residues811. In an attempt to overcome some of these limitations, we recently reported peptide ubiquitylation with the temporary ligation auxiliary, 2-aminooxyethanethiol, which employed reduction with metallic Zn in the terminal step12,13. Unfortunately, we found that efcient NO bond reduction required both harsh denaturants and strongly acidic conditions. Therefore, our approach was ultimately limited to peptides or proteins amenable to refolding from the denatured state.

Herein we report the discovery of an unprecedented aromatic thiol-mediated NO bond cleavage reaction that is compatible with folded proteins at physiological pH and that overcomes limitations of current strategies for chemical ubiquitylation. Mechanistic investigation of this new reaction implicates a disulde radical anion as the reductive species that cleaves NO bonds. The semisynthesis of full-length human histone H2B modied by the small ubiquitin-like modier protein, SUMO-3, demonstrates the complete compatibility of this reaction with thiol side-chains in folded proteins and signicantly expands the practical scope of chemical ubiquitylation.

Results4-Mercaptophenylacetic acid-mediated NO bond cleavage. We previously reported the successful application of the auxiliary 2-aminooxyethanethiol towards peptide ubiquitylation13. The utility of this auxiliary group lies in its high-yielding 3-step synthesis and easy incorporation in various peptide substrates. However, two challenges in removing the auxiliary and producing a wild-type amide linkage were the requirement for pH 3, and the necessity of chaotropes such as 6 M guanidinium chloride that unfold ubiquitin and allow reduction of the NO bond by metallic Zinc (Fig. 1). Although such a strategy is compatible with proteins that may be refolded from the denatured state, its broad utility is limited. Moreover, the electrophilic character of the nascent disubstituted amide bond in the ligation product led to a small amount of hydrolysis over time, which was exacerbated at the low pH required for efcient NO bond reduction.

In an effort to reduce the amount of hydrolysed ubiquitin (175)-COOH side-product and to increase the rate of transthioesterication between auxiliary-bearing peptides and the ubiquitin(175)-a-thioester, we tested the aromatic thiol 4-mercaptophenylacetic acid (MPAA) as a ligation additive (Fig. 1). The excellent leaving group ability of MPAA renders its protein thioesters more reactive towards transthioesterication, the rst and rate-limiting step in native

chemical ligation14,15. To our surprise, in a typical ligation reaction with 0.5 mM ubiquitin(175)-a-thioester and 5 mM of auxiliary-bearing peptide (KAKauxI) in a buffer consisting of 50 mM Tris, 150 mM NaCl and 200 mM MPAA at pH 7.3, we observed the nal ligation product to be altogether missing the ligation auxiliary (Fig. 2a and Supplementary Figs 14). This unexpected result was consistently reproducible, although the slow kinetics of product formation necessitated up to 48 h to achieve 5070% yields (Fig. 2b). Additional controls revealed MPAA to be the critical component required for NO bond cleavage, and re-purication of the commercial compound by high-performance liquid chromatography (HPLC) did not inhibit the reaction (Supplementary Table 1, entries 18 and Supplementary Fig. 5). The necessity of a free SH group in MPAA was seen from the fact that pure disulde-linked MPAA dimer did not undertake NO bond cleavage (Supplementary Fig. 6 and Supplementary Table 1, entry 9).

NO bond reduction by aromatic and aliphatic thiols. In order to ascertain the generalizability of this unprecedented NO bond cleavage reaction, we tested a range of aromatic and aliphatic thiols with the ubiquitylated ligation product KAKUb(aux)I, bearing the auxiliary at the site of ligation (Supplementary Fig. 7). With the exception of 4-hydroxythiophenol, all aromatic thiols tested undertook NO bond cleavage (Table 1, entries 15). Interestingly, 4-hydroxythiophenol inhibited MPAA-mediated NO bond cleavage when equal amounts of both were present in the reaction mixture. In contrast, none of the aliphatic thiols led to NO bond reduction under identical buffer conditions (Table 1, entries 610). This suggests that near a neutral pH the NO bond is compatible with aliphatic thiol additives commonly employed in native chemical ligations, such as 2-mercaptoethanesulfonic acid (MESNa), and that it is also stable to the biological reducing agent glutathione.

One key difference between the two compound classes tested is that the aromatic thiols are 450% deprotonated at pH 7.3, while the aliphatic thiols are largely protonated. The importance of a deprotonated thiolate species was suggested by the fact that NO

CONH2

NH

O O

O

n

e

-

p

o

t

H2N

HN O

N

H N

H

HS

O

COOH MPAA

H2N

L

i

g

a

t

i

o

n

Figure 1 | Aromatic thiol-mediated one-pot traceless native chemical ubiquitylation. MPAA, 4-mercaptophenylacetic acid; UbDG76-SR, ubiquitin(175)-a-thioester with 2-mercaptoethanesulfonic acid. PDB code 1UBQ (ubiquitin).

2 NATURE COMMUNICATIONS | 7:12979 | DOI: 10.1038/ncomms12979 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12979 ARTICLE

a b

0.9

0.8

0.6

0.5

0.3

0.2

+6

Intensity

Fraction product

Degassed300 l reaction

1.5 ml headspace 300 l reaction

1.7 ml headspace 100 l reaction

1.7 ml headspace 100 l reaction

50 mM H2O2

+7

+5

+8 +9

+10

800 1,200 1,600 2,000

Mass (m/z)

0 4 8 12 16 20 24 28 32 36 40 44 48

Time (h)

Figure 2 | NO bond cleavage in the native chemical ligation product KAKUb(aux)I. (a) ESI-MS spectrum of the nal ligation product of ubiquitin(175)-athioester with KAKauxI. Calculated for KAKUb(aux)I, 9,081.1 Da. Observed KAKUbI, 9,004.82.7 Da. (b) Time-course of NO bond cleavage and KAKUbI formation from the auxiliary-containing test substrate KAKUb(aux)I in a buffer consisting of 200 mM MPAA, 100 mM NaH2PO4 at pH 7.3 under the indicated conditions. Error bars represent the s.d. from three independent measurements.

Table 1 | Aromatic and aliphatic thiol pKa values and corresponding yields in NO bond cleavage assays.

Entry Thiol pKa14,29 Yield (%)

1 4-Nitrothiophenol 4.5 59 2 3-Mercaptobenzoic acid 5.8 81 3 4-Mercaptophenylacetic acid 6.6 73 4 4-Aminothiophenol 6.9 73 5 4-Hydroxythiophenol 7.0 n.d. 6 2,2,2-Triuoroethanethiol 7.6 n.d.7 L-Glutathione 9.1 n.d.8 D,L-Dithiothreitol 9.2, 10.1 n.d. 9 2-Mercaptoethanesulfonic acid 9.2 n.d. 10 2-Methyl-2-propanethiol 11.2 n.d.

n.d. no detectable NO bond cleavage.

bond cleavage by MPAA was dramatically reduced at pH 6.0 (Supplementary Table 1, entries 1012).

Mechanistic investigation of the NO bond cleavage reaction. Examples of bioactive compounds with chemically labile NO bonds include pro-drug forms of the duocarmycin and CC-1065 class of antitumor agents16. These were proposed to undergo conversion to a biologically active form upon cleavage of the NO bond by nucleophilic thiols within the tumor microenvironment. Since we observed an increase in NO bond cleavage with increasing pH, a nucleophilic mechanism may in principle also be invoked for the aromatic thiols. However, freeze-thaw degassing the reaction mixtures under an Argon atmosphere sufced to inhibit the reaction with MPAA at pH 7.38.5 (Fig. 2b and Supplementary Table 1, entry 1314), ruling out a purely nucleophilic mechanism and suggesting a key role for dissolved oxygen. In support of the latter, increasing the reaction headspace, and thereby the ratio of molecular oxygen to thiol, resulted in halving the reaction time to 24 h (Fig. 2b).

Molecular oxygen may act as a terminal electron acceptor and favour the formation of aromatic thiyl radicals from aromatic thiolates17. One indication that thiyl radicals may be present under the reaction conditions was our observation of the oxidized MPAA disulde species. To test the possibility that aromatic thiyl

radicals are spontaneously formed in buffered aqueous solutions at pH 7.3, we incubated a range of aromatic thiols with the co-factor nicotinamide adenine dinucleotide (NADH) (Supplementary Fig. 8). Thiyl radicals react with NADH to yield the NAD radical and the consumption of NADH is readily detected by a decrease in absorbance at 340 nm (ref. 18).

In buffers that favoured NO bond reduction, we also observed a dramatic decrease in NADH concentration. Importantly, and consistent with their inability to reduce the NO bond, we did not observe similar oxidation of NADH in reactions with aliphatic thiols 610 in Table 1 over a 6 h time-course. The disparity in reducing nature of aromatic and aliphatic thiols was further seen by their reaction with methyl viologen (MV2). A rapid increase in absorption at 605 nm characteristic of the single-electron transfer reduction product MV was observed only with aromatic thiols, suggesting the formation of a strongly reducing species in solution19.

Electron paramagnetic resonance (EPR) is a widely employed technique to detect species with unpaired electrons. However, thiyl radicals are generally not directly detectable by EPR in solution due to the large spin-orbit coupling constant of sulfur, which leads to fast relaxation of the electron spin20. Hence we attempted indirect EPR detection of the MPAA radical in solution at room temperature by spin-trapping with the compound 5,5-dimethyl-1-pyrroline-N-oxide (DMPO)21. To our delight, we observed the formation of nitroxide radicals in a solution containing 50 mM MPAA and 100 mM DMPO dissolved in 50% (v/v) aqueous N,N-dimethylformamide (DMF) (Fig. 3a). This was attributed to the addition of a thiyl radical into DMPO. The spectrum exhibited six lines characteristic of DMPO-trapped thiyl radicals22, which could be simulated with aN 14.22 G and aH 16.16 G (Fig. 3b), giving aNoaH as

expected for DMPO-thiyl radical adducts in aqueous solution23,24. The inclusion of 50 mM Na2HPO4, pH 7.5 in the reaction altered the appearance of the EPR spectrum (Fig. 3c), but controls lacking MPAA did not show EPR signal under these conditions (Supplementary Fig. 9). Furthermore, trapping of the MPAA thiyl radical was conrmed by electrospray ionisation mass spectrometry (ESI-MS) (Fig. 3d and Supplementary Fig. 10). Importantly, alkylation of MPAA in situ with 70 mM 2-iodoacetamide for 1.5 h precluded the appearance of an EPR signal, providing further evidence for a thiyl radical as the reactive species (Fig. 3e). In support of MPAA thiyl radical formation in

NATURE COMMUNICATIONS | 7:12979 | DOI: 10.1038/ncomms12979 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12979

a

1:1 DMF:H2O

50 mM MPAA, 100 mM DMPO

1:1 DMF:H2O, 50 mM Na2HPO4 50 mM MPAA, 100 mM DMPO

1:1 DMF:H2O, 50 mM Na2HPO450 mM MPAA + 70 mM 2-iodoacetamide, 100 mM DMPO

10 G

b

aN = 14.22 G aH = 16.16 G

c

d

O

N S O

OH

Intensity

200 300 400 Mass (m/z)

281.3

[M+H+] = 281.1 Da

e

Figure 3 | EPR spectra of DMPO/ S-Ar adduct. (a) Spectrum obtained

upon incubating 50 mM MPAA and 100 mM DMPO in a 1:1 water-DMF mixture at 25 C. (b) Computer simulation of the spectrum observed in a with hyperne splitting constants aN 14.22 G and aH 16.16 G23.

(c) Spectrum obtained upon incubating 50 mM MPAA and 100 mM DMPO in 50 mM Na2HPO4 at pH 7.5, in a 1:1 water-DMF mixture at 25 C.

(d) ESI-MS spectrum obtained by LC-ESI-MS analysis of the reaction components in c. Inset shows the proposed radical adduct. (e) EPR spectrum obtained upon pre-incubating 50 mM MPAA with 70 mM 2-iodoacetamide for 1.5 h followed by 100 mM DMPO in 50 mM Na2HPO4 at pH 7.5, in a 1:1 water-DMF mixture at 25 C. Incubation of 50 mM MPAA and 100 mM DMPO in 50 mM Na2HPO4 at pH 7.5, in a 1:1 water-DMF mixture at 25 C for 1.5 h without the addition of 2-iodoacetamide resulted in a spectrum similar to that seen in c. Spectrometer settings: microwave power, 20 mW; modulation amplitude, 1.0 G; time constant, 163 ms;scan rate, 0.6 G/s.

the presence of 50 mM Na2HPO4, we recapitulated the spectrum observed in Fig. 3c by adding a small volume of 600 mM Na2HPO4 to a nal concentration of 50 mM in the 50%

(v/v) aqueous DMF mixture (Supplementary Fig. 11). Next, we included 50 mM H2O2 in the solution containing 50 mM MPAA and 100 mM DMPO dissolved in 50% (v/v) aqueous DMF. We

observed a spectrum identical to that without H2O2 present, yet with much greater signal intensity, and again observed the mass of the DMPO-MPAA adduct (Supplementary Fig. 12a). However, we did not see any signicant nitroxide radicals forming in the absence of MPAA, and any DMPO-OH adduct expected to arise from H2O2 alone was only observable by ESI-MS (Supplementary

Fig. 12b). Finally, if the MPAA radical is crucial for NO bond cleavage, we expected its trapping by DMPO to inhibit the reaction. Indeed, we found that the addition of 1 M DMPO to a reaction containing 100 mM MPAA inhibited product formation over the course of 24 h.

One potential pathway for radical-mediated NO bond cleavage is by formation of a thiyl radical in the auxiliary (Supplementary Fig. 13). A 1,3-sigmatropic rearrangement of the thiyl radical would result in a carbon-centered radical adjacent to the low energy NO bond, which favours its homolysis25. In order to test this mechanism we alkylated the auxiliary thiol in the ligation product with N-(2-chloroethyl)-N,N-dimethylammonium chloride (Supplementary Fig. 14), thereby precluding formation of a thiyl radical. Upon treatment with 200 mM MPAA at pH 7.3 we still observed efcient NO bond cleavage in the S-alkylated product, indicating that a substrate-derived thiyl radical is not essential for the reaction to proceed. An alternative pathway for NO bond cleavage is direct reduction by a reducing agent. In thinking of reducing species that are generated by a combination of aromatic thiolates and thiyl radicals, we considered the possibility of a disulde radical anion. The formation of this high-energy species has been observed with small molecule thiols such as cysteine and glutathione26, and in the active site of the enzyme ribonucleotide reductase27. The presence of an unpaired electron in an antibonding s* orbital renders the disulde radical anion a strongly reducing yet transient species, in equilibrium with the dissociated radical and thiolate forms28.

We surmised that slow formation of the thiyl radical by molecular oxygen in the absence of added radical initiators along with the transient nature of the disulde radical anion together contribute to the slow kinetics of NO bond cleavage. An initial attempt to increase the rate of MPAA-mediated NO bond cleavage by including the water-soluble radical initiator, 2,2-azobis[2-(2-imidazolin-2-yl)propane]dihydrochloride (VA-044) in our reactions proved unfruitful. This is likely due to the rapid quenching of the carbon-centered radical before the formation of signicant amounts of the disulde radical anion and is consistent with the reported inhibition of radical-mediated desulfurization of cysteine residues by MPAA, which also employs VA-044 (ref. 29).

We also considered if superoxide, formed en route to the thiyl radical, may act as a reducing agent30. To test this possibility, we utilized the well established superoxide generating system consisting of xanthine oxidase and its substrate hypoxanthine31. xanthine oxidase catalyses the conversion of hypoxanthine rst to Xanthine then to uric acid, and superoxide is released at each step. We performed this reaction in the presence of KAKUb(aux)I, with hypoxanthine at 4-fold excess relative to the auxiliary containing test substrate. By monitoring the appearance of uric acid at 290 nm, we observed that all of the hypoxanthine was converted to uric acid within the rst minute of reaction, producing a burst of superoxide (Supplementary Fig. 15). However, even after 24 h no cleavage of the auxiliary was observed, strengthening our hypothesis that a disulde radical anion is the likely reductant (Supplementary Table 2, entry 1).

Known methods to generate disulde radical anions from thiols or disuldes in solution include ash photolysis28, pulse radiolysis32 or cyclic voltammetry33, all of which are technically challenging in the presence of folded protein substrates.

4 NATURE COMMUNICATIONS | 7:12979 | DOI: 10.1038/ncomms12979 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12979 ARTICLE

Therefore, we wondered if a mild oxidant, such as hydrogen peroxide (H2O2), would facilitate the formation of thiyl radicals by Fenton chemistry (Fig. 4a)34,35. Indeed, the inclusion of 50 mM H2O2 with 200 mM MPAA led to a signicant increase in the rate of product formation, requiring only 4 h to attain maximal conversion with no detectable amounts of undesired protein oxidation (Fig. 2b and Supplementary Table 2, entry 2). The chelation of free metal ions with 50 mM EDTA effectively inhibited the reaction, indicating the key role for trace metal ions in generating thiyl radicals (Supplementary Table 2, entry 3). Since the concentration of trace metal ions in reaction components may vary, we demonstrated that 1 mM FeCl2 may be added to facilitate the reaction with no deleterious effect on reaction yield (Supplementary Table 2, entry 4). Importantly, 50 mM H2O2 alone or mixed with 1 mM FeCl2 did not yield detectable product in the absence of MPAA, proving that Fenton chemistry alone cannot undertake NO bond cleavage (Supplementary Table 2, entries 56). Finally, freeze-thaw degassing a solution of 50 mM H2O2 and 200 mM

MPAA failed to prevent reductive chemistry, which conrmed our hypothesis that H2O2 can favour thiyl radical formation even in the absence of molecular oxygen (Fig. 4a and Supplementary Table 2, entry 7).

Computational studies of NO bond cleavage. We next undertook ab initio quantum chemistry calculations to interrogate the feasibility of disulde radical anion formation and its reactivity towards the NO bond. The relative redox potentials (DE) for (1) electron transfer between various thiols and hydrogen peroxide (Fig. 4b), and (2) subsequent electron transfer

between the disulde radical anion and a model diglycine compound, 1 (Fig. 4c), were computed. DE values for the rst step were obtained by calculating the free energy changes of the redox reactions using the Gaussian 09 program package36. Equilibrium geometries of all species were located via geometry optimization and the thermal corrections were evaluated at the B3LYP/6-31G* level of theory, with the solvent effect modelled using the polarizable continuum model. The electronic energies at the equilibrium geometries were computed at the B3LYP/6-311 G** level of theory and the

results are summarized in Table 2.

Our calculations revealed that protonated aliphatic thiols do not favour disulde radical anion formation (Table 2, entries 58). In contrast, deprotonated aromatic thiolates can be oxidized by hydrogen peroxide to form disulde radical anions (Table 2, entries 14). For MPAA, the calculated DE of 2.07 V (Table 2, entry 3) indicates that disulde radical anion formation

Table 2 | Calculated redox potentials for disulde radical anion formation from aliphatic and aromatic thiols at pH 7.3.

Entry Thiol DE (V) 1 4-Nitrothiophenol 1.58 2 3-Mercaptobenzoic acid 1.98 3 4-Mercaptophenylacetic acid 2.07 4 4-Aminothiophenol 2.035 L-Glutathione 0.40

6 D,L-Dithiothreitol 0.33

7 2-Mercaptoethanesulfonic acid 0.43

8 2-Methyl-2-propanethiol 0.83

a b

H2O2 + Fe2+

H2O2 + Fe3+

OH + ArSH

HOO + ArSH

OH + ArSH

ArS + ArS

H N

H N

(1)

(2)

(3)

(4)

(5)

(6)

Fe3+ + OH + OH 4 ArS + H2O2 + 2H+

4 RSH + H2O2RSH = L-Glutathione, D,L-Dithiothreitol,

2-Mercaptoethanesulfonic acid, 2-Methyl-2-propanethiol

2 ArS

SAr + 2 H2O

Eo= 1.58 to 2.07 V

Eo= 0.83 to 0.33 V

Fe2+ + HOO + H+

Ar= p-(NO2)C6H4, m-(CO2H)C6H4, p-(CH2CO2H)C6H4, p-(NH2)C6H4,

ArS + H2O

ArS + H2O2

ArS + H2O

ArSS Ar

2 RS

SR + 2 H+ + 2 H2O

c

O

O O

(ArS)2

H N

H N

H N

H N

H N

N

N N

N

O

O O O O

O

O

O

O

(ArS)2

O

SH SH

SH

HS

H N

O

H N

H N

N

H

O

O O O O

1

Figure 4 | Formation of disulde radical anions and their role in NO bond cleavage. (a) Production of aromatic thiyl radicals mediated by trace-metal-catalysed Fenton chemistry (14) and their combination with aromatic thiolates to form disulde radical anions (56). (b) Net chemical equations for the formation of disulde radical anions from aromatic thiolates and aliphatic thiols at pH 7.3. The calculated range of standard redox potentials is indicated for compounds from each class of molecules. (c) Proposed mechanism for disulde radical anion-mediated NO bond cleavage in the model compound 1.

NATURE COMMUNICATIONS | 7:12979 | DOI: 10.1038/ncomms12979 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12979

is thermodynamically favoured. However, the rate of formation is not predictable by computation and disulde radical anions are known to exist in an equilibrium that favours the dissociated radical and thiolate species27, which likely underlies the slow kinetics of reduction. Next, we focused on electron transfer from the MPAA disulde radical anion to 1 (Fig. 4c). Based on the calculated electron densities in the highest occupied molecular orbital of the disulde radical anion and the lowest unoccupied molecular orbital in the diglycine peptide (Supplementary Fig. 16) two possible pathways exist. We found both pathways to be thermodynamically permissible, with DE of 0.53 V for the production of a b-mercaptoethyloxyl radical, and DE of 0.04 V for the production of the b-mercaptoethanolate anion. When up to nine explicit H2O molecules in varying combinations were included in these calculations, most instances within the test set favoured the b-mercaptoethanolate anion (Supplementary

Table 4). Finally, we also examined the thermodynamics of direct electron transfer from the MPAA thiolate to 1 and found this process to be energetically unfavourable, which suggests that the thiolate form alone cannot act as a reductant.

Mechanistic studies with a model diglycine compound. In order to unambiguously identify the b-mercaptoethanol predicted by our proposed mechanism (Fig. 4c), we synthesized the S-trityl-protected form of the model dipeptide, compound 2 (Supplementary Figs 17 and 18). Compound 2 was sparingly soluble in water and hence subjected to reduction with 200 mM MPAA in a buffer consisting of 100 mM Na2HPO4, pH 7.3 in 50% (v/v) aqueous DMF. Consistent with results obtained from protein substrates, and the detection of a MPAA radical under these conditions, we observed cleavage of the NO bond over 24 h. The S-trityl-protected b-mercaptoethanol was isolated and conrmed by NMR (Supplementary Fig. 19). Surprisingly, in the presence of 50 mM H2O2 and 200 mM

MPAA, complete NO bond cleavage in 20 mM of 2 was observed in 10 min. This may reect the greater accessibility of the labile bond in 2 than in ubiquitylated peptides, or more productive electron transfer arising from a smaller number of competing amide bonds than in ubiquitin. That trityl protection of the auxiliary thiol did not prevent NO bond cleavage underscores the fact that electron transfer to the NO bond occurs directly from the reducing species. As expected, sequestration of trace metal ions by treatment of the buffer with a metal-chelating resin also inhibited NO bond cleavage in the model compound (Supplementary Table 3).

One-pot strategy for native chemical ubiquitylation. Current chemical ubiquitylation methodologies are not optimal for native folded proteins29,37. Therefore, an efcient one-pot method to perform native ubiquitylation is highly desirable. With this in mind, we sought to improve the yield of one-pot auxiliary-mediated ubiquitylation by limiting the formation of ubiquitin (175)-COOH. Slow hydrolysis of both the ubiquitin(175)-athioester and the disubstituted amide in the ligation product contribute to this undesired side-product, which can in principle be alleviated by enhancing the kinetics of both ligation and NO bond reduction. Surprisingly, we found that conducting the ligation reaction without MPAA signicantly avoided ubiquitin(175)-COOH formation by preventing premature auxiliary removal in the starting materials. The subsequent addition of 200 mM MPAA to the crude ligation mixture and incubation for an additional 24 h at 25 C generated the nal reduced ubiquitylated peptide in 82% overall yield. This represents a 20% higher yield over reactions where MPAA was added at the start of ligation.

Synthesis of full-length sumoylated histone H4. As an initial test of the auxiliarys utility in the context of folded proteins, we incorporated non-denaturing MPAA-mediated auxiliary removal into the semisynthesis of full-length sumoylated human histone H4 (suH4). Although histone sumoylation was rst reported over a decade ago, very little is known regarding its functional role in human chromatin38. Access to quantities of H4 site-specically conjugated with the C terminus of SUMO-3 at Lys12 is crucial for biochemical investigations of the role for sumoylation in regulating chromatin structure and function39. We devised a synthetic strategy for suH4 (Supplementary Fig. 20 and Supplementary Methods). The H4 (114) peptidyl hydrazide was synthesized using Fmoc chemistry on the solid phase, with Gly92 of SUMO-3 and the ligation auxiliary attached to Lys12 (Supplementary Fig. 21). Following release from the solid-phase and global deprotection, the peptide was ligated to the SUMO-3(291)C47S-a-thioester to yield the sumoylated peptide (Supplementary Figs 22 and 23). One key challenge we anticipated was conversion of the hydrazide to a thioester without cleaving the NO bond. However, we found the auxiliary to be completely stable to both diazotization and thioester formation, which employed NaNO2 at pH 3.0 followed by displacement of the resulting azide with 100 mM MPAA40. Complete retention of the auxiliary through these steps highlights its utility in diverse native chemical ligation strategies. The sumoylated H4 peptide a-thioester was then reacted with a truncated H4(15102) protein containing the A15C mutation at its N terminus to facilitate native chemical ligation (Supplementary Fig. 24). Ligation proceeded over 24 h to afford 2.1 mg of the ligated product, retaining the ligation auxiliary, in 66% puried yield (Supplementary Fig. 25). The ligation product was then dissolved in a buffer consisting of 100 mM Na2HPO4, 200 mM MPAA, pH 7.3 and NO bond cleavage allowed at 25 C over 24 h to yield the reduced compound (Supplementary Fig. 26). Importantly, Cys15 in H4 was unaffected by MPAA-mediated auxiliary removal, demonstrating the compatibility of this reaction with folded proteins containing Cys residues. In the terminal step, the full-length sumoylated histone H4 A15C mutant was desulfurized to yield the desired suH4 in 41% yield over the last two steps (Supplementary Fig. 27).

Synthesis of full-length sumoylated histone H2B. The ultimate goal for our chemical strategy is complete compatibility with native folded proteins containing Cys residues. We envision future applications wherein a suitably protected ligation auxiliary is directly incorporated in target proteins by employing an amber suppression strategy and puried from producer strains before native ubiquitylation/sumoylation41. Therefore, having demonstrated that MPAA can mediate auxiliary removal from sumoylated histone H4 under non-denaturing conditions, we sought to perform (1) auxiliary deprotection, (2) sumoylation and(3) the auxiliary removal step on an additional protein target without intermediate denaturation and purication steps. We were particularly attracted to histone H2B as it is sumoylated at its C-terminal Lys120 (suH2B; ref. 42) and genetic experiments suggest that sumoylation recapitulates the genomic occupancy of H2B Lys120 ubiquitylation43. However, similar to suH4, the role of suH2B in chromatin regulation awaits in vitro biochemical investigation.

Towards the semisynthesis of suH2B, we rst generated full-length histone H2B bearing a protected ligation auxiliary as the entry point for testing our methodology. To ensure that the auxiliary protecting group could be removed under native conditions, we synthesized a photoprotected form (3) by starting from 2-nitrobenzyl chloride and N-(2-bromoethoxy)phthalimide (Fig. 5a and Supplementary Figs 2830)13. The auxiliary 3 was

6 NATURE COMMUNICATIONS | 7:12979 | DOI: 10.1038/ncomms12979 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12979 ARTICLE

a

b

NO2

i

S SH

O

O

NO2 NO2

ii iii

N O

H2N

NO2 NO2

NO2

3

O

S iv S

S

O

Cl

O

O

i

O

O SH

SH

ii

O BocHN

TrtS

H2N OH

( )

3

HS

( )

3

O

NH

SR

( )

3

HN

O

NH

ivDde

S N H

HN

O

4

O

O

H2B(117125)A117C

SUMO-3(291)C47S-MES

H2B(1116)-MES

NO2

5

iii iv

SH

SH SH

HN

O

HN O

O

( )3 ( )3

SH v

O

O

7 8

O

N

( )3

HN

O

SR

HN HN

O

6

Figure 5 | Semisynthesis of full-length sumoylated histone H2B(A117C). (a) Synthesis of photoprotected auxiliary 3. (i) CH3C(O)SH, K2CO3, THF, 8 h, 25 C. (ii) HCl, CH3OH, 6 h, 60 C, 75% (2 steps). (iii) N-(2-bromoethoxy)phthalimide, Et3N, DMSO, 4 h, 25 C, 74%. (iv) H2NNH2, CHCl3, 1 h, 25 C, 98%.

(b) (i) Site-specic coupling of 3 to H2B(117125)A117C Lys120 followed by acidolytic release of the unprotected peptide, 4, from the solid-phase.(ii) Expressed protein ligation of 4 with H2B(1116)-a-thioester to generate full-length H2B(A117C) with protected auxiliary at Lys120, 5. (iii) photolytic removal of the auxiliary protecting group to give H2B(A117C) with unprotected auxiliary at Lys120, 6. (iv) Expressed protein ligation of 6 with SUMO-3 (291)C47S-a-thioester to generate sumoylated H2B(A117C) 7, with retention of the ligation auxiliary. (v) Selective removal of the ligation auxiliary with 150 mM MPAA under non-denaturing conditions to yield sumoylated H2B(A117C) 8. ivDde 1-(4,4-Dimethyl-2,6-dioxocyclohexylidene)-3-methylbutyl

group. PDB codes, 1KX5 (H2B) and 1U4A (SUMO-3).

incorporated at Lys120 of the H2B(117125) C-terminal peptide with an Ala to Cys mutation at position 117 (Fig. 5b). After acidolytic cleavage from the solid phase, the peptide 4 was ligated via its N-terminal Cys to an H2B(1116)-a-thioester to generate full-length H2B(A117C), 5 (Fig. 6a,b and Supplementary Figs 3133). The product 5 was folded by dialysis into 50 mM Na2HPO4, pH 7.5 (Fig. 6c), and every subsequent step was performed under folded conditions. First, complete deprotection of the auxiliary thiol was achieved by irradiation with 365 nm light for 3.5 h in the presence of ascorbic acid, semicarbazide and dithiothreitol. The integrity and folded state of the deprotected protein were conrmed by ESI-MS and circular dichroism (Fig. 6d,e). The deprotected H2B(A117C)aux, 6, was then ligated under non-denaturing conditions to the SUMO-3 (291)C47S-a-thioester over 48 h to yield the ligation product

H2B(A117C)Su(C47S)aux, 7 (Fig. 6fh and Supplementary Fig. 34). The sumoylated product 7 was subjected to MPAA-mediated auxiliary removal for 24 h, yielding the ligation product lacking the ligation auxiliary, H2B(A117C)Su(C47S), 8, in 1530% yield over two steps (Fig. 6i). Importantly, we observed no precipitation of the H2B species throughout these manipulations, and the folded state of the reduced ligation product 8 was conrmed by size exclusion chromatography and circular dichroism (Supplementary Fig. 35). As a functional test of the correct folded state of SUMO-3 in 8, we undertook its desumoylation with the SUMO-specic protease sentrin specic peptidase 1 (SENP1). Congruent with our observations that MPAA-mediated NO bond cleavage did not lead to the denaturation or aggregation of 8, we observed efcient hydrolysis by SENP1 and the appearance of lower molecular weight species corresponding to the hydrolysed SUMO-3 and H2B(A117C) (Supplementary Fig. 36).

DiscussionThe NO bond is frequently encountered in organic chemistry and several methods exist for the cleavage of this moiety, including TiCl3 (ref. 44), catalytic hydrogenation45, Na/Hg

amalgams46 and SmI2 (ref. 47). More recently, neutral organic super-electron donors were demonstrated to reduce NO bonds in Weinreb amides48. However, the application of any of these reagents to folded proteins in aqueous buffers is extremely challenging. Our discovery of an unprecedented NO bond reductive chemistry sets the stage for new applications that would benet from the controlled reversal of this low-energy bond. The straightforward synthesis of the 2-aminooxyethanethiol auxiliary and its photoprotected form, their facile incorporation into peptides, and cleavage by a subset of readily available water-soluble aromatic thiols is particularly appealing for applications in protein semisynthesis. Both experimental and computational investigations support our hypothesis that NO bond cleavage may involve the formation of a transient disulde radical anion species. Multiple observations towards this include the requirement for an oxidant, the necessity for trace metal ions, and the detection of aromatic thiyl radicals. Importantly, our ability to readily control the timing of NO bond cleavage is particularly appealing as demonstrated by the one-pot synthesis of the ubiquitylated peptide, KAKUbI. As highlighted in our syntheses of the full-length sumoylated human histones H4 and H2B, the extremely mild reductive strategy may also be applied towards the sumoylation, and by extension ubiquitylation, of native folded proteins in aqueous buffers. Indeed, the semisynthesis of suH4 and suH2B will, for the rst time, permit detailed biochemical studies of these poorly understood modications. Finally, having established the stability of the ligation auxiliary to the intracellular reductant glutathione, our immediate future efforts are focused on an amber-codon-suppression strategy to incorporate the auxiliary into natively folded proteins that are inaccessible by fragment-based semisynthetic approaches.

Methods

Solid-phase peptide synthesis. All peptides were synthesized by standard 9-uorenylmethoxycarbonyl (Fmoc)-based solid-phase peptide synthesis on a Liberty Blue Automated Microwave Peptide Synthesizer (CEM, Matthews, NC). The auxiliary containing peptides KAKauxI, H4(114)aux-C(O)NHNH2, and

NATURE COMMUNICATIONS | 7:12979 | DOI: 10.1038/ncomms12979 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12979

H2B(117125, A117C)photoaux-C(O)OH were synthesized on Rink-amide, 2-chlorotrityl chloride, and Wang resin, respectively. The ligation auxiliary was incorporated by coupling bromoacetic acid at the desired Lys e-amine followed by on-resin displacement of the bromide with 0.25 M protected auxiliary, O-(2-(tritylthio)ethyl) hydroxylamine or O-(2-((2-nitrobenzyl)thio)ethyl)hydroxylamine, in DMSO.

One-pot ligation and auxiliary removal. The ubiquitin(175)-a-thioester (1 eq.) and KAKauxI (10 eq.) were dissolved in a reaction buffer containing 50 mM tris, 150 mM NaCl, 10 mM tris(2-carboxyethyl)phosphine (TCEP) pH 7.3, and incubated at 25 C for 24 h. Following ligation, MPAA was added to a nal

concentration of 200 mM, and the reaction subsequently incubated at 25 C for a further 24 h.

General method for NO bond cleavage. Auxiliary-containing protein substrates were dissolved in 200 mM MPAA, 100 mM Na2HPO4, pH 7.3 at a concentration of0.1 mM. Reactions were either performed in a reaction vessel with headspace at least 15 times that of the reaction volume at 25 C for 24 h, or supplemented with 50 mM H2O2 and incubated at 25 C for 4 h.

Analysis of auxiliary removal from protein substrates. Reaction mixtures were treated with 50 mM TCEP, pH 7.3, at 4 C for 30 min, then acidied to pH B3 with formic acid. Samples were extracted once with diethyl ether to remove a majority of the aromatic thiol and then analysed by C18 liquid chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS) employing a gradient of 5100% acetonitrile over 40 min.

Synthesis of sumyolated H4. Puried SUMO-3(291)C47S-MESNa thioester(1 eq.) and H4(114)aux-C(O)NHNH2 (6 eq.) were reacted in a buffer containing6 M Gn-HCl, 100 mM Na2HPO4, and 10 mM TCEP, pH 7.3. Ligation proceeded with gentle shaking at 25 C for 24 h. The ligation product, H4(114)Su(C47S)aux-C(O)NHNH2, was puried by C18 preparative reversed-phase (RP)-HPLC, then converted to an acyl azide by reaction with 15 eq. of NaNO2 in 200 mM Na2HPO4,6 M Gn-HCl, pH 3.0, at 20 C for 15 min. A solution of H4(15102)A15C (2 eq.)

in 200 mM Na2HPO4, 6 M Gn-HCl, 200 mM MPAA, pH 6.5, was then added to the thioester and the mixture was allowed to warm up to room temperature. The pH was adjusted to 6.87.0 and ligation allowed at 25 C for 24 h. The ligation product was puried by C4 semi-preparative RP-HPLC, dissolved in 200 mM MPAA, 100 mM Na2HPO4, pH 7.3, and incubated at 25 C for 24 h to remove the auxiliary.

Product lacking the auxiliary group was further puried by C4 semi-preparative RP-HPLC and subjected to desulfurization in 280 mM 2-methyl-2-propanethiol, 10 mM VA-044, 100 mM Na2HPO4, 6 M Gn-HCl, 500 mM TCEP, 100 mM

MESNa, pH 7.5. The reaction proceeded at 37 C for 24 h, and the nal desired product was puried by C4 analytical RP-HPLC.

Synthesis of sumoylated H2B. Puried H2B(1116)-MESNa thioester (1 eq.) and H2B(117125, A117C)photoaux-C(O)OH (10 eq.) were reacted in a buffer containing 6 M Gn-HCl, 100 mM Na2HPO4, 10 mM EDTA and 5 mM TCEP,pH 7.5. Ligation proceeded with gentle shaking at 25 C for 6 h. The ligation product, H2B(A117C)photoaux, was puried by C4 preparative RP-HPLC, then folded at 0.25 mg ml 1 by dialysis at 4 C from 6 M Gn-HCl, 100 mM Na2HPO4, pH 7.5 into 50 mM Na2HPO4, pH 7.5. Unmasking of the auxiliary thiol was accomplished by adjusting the buffer composition to 50 mM Na2HPO4, 4 mM semicarbazide, 5 mM ascorbic acid, 0.5 mM dithiothreitol, pH 67, and irradiating with 365 nm light for 3.5 h. Deprotected H2B(A117C)aux was dialysed back into 50 mM Na2HPO4, pH 7.5, and to this solution was added SUMO-3(291)C47S

MESNa thioester (3 eq.) dissolved in 50 mM Na2HPO4, pH 7.5. A solution containing 200 mM TCEP, 20 mM MPAA, 50 mM Na2HPO4, pH 7.5 was added to the reaction to attain nal concentrations of 2 mM TCEP and 0.2 mM MPAA. The ligation proceeded at 22 C for 48 h. Ligation was conrmed by LC-ESI-MS and SDSpolyacrylamide gel electrophoresis analysis. MPAA was then added to the reaction to a nal concentration of 150 mM and the reaction transferred to a container with headspace lled with air equal to 10 times the liquid volume of the reaction. MPAA-mediated auxiliary removal proceeded for 24 h at 22 C, and the

a

NO2

SH SH

S

SH

( )3

hv

HN HN

HN HN

O O

( )

3

O

O

5 6

b c

5 5

6

15

15

Intensity 10 3 (deg cm 2

dmol1residue1)

+14

+12

+10

Intensity Intensity

+16

+18

600

195 250

[afii9838] (nm)

15

+8

1,800

Mass (m/z)

d e

6

+14

+16 +12

+10 +18

600 1,800

O N

O

103(degcm2

dmol1residue1)

+8

195 250

[afii9838] (nm)

15

Mass (m/z)

f

SH

SH

O

( )

3 ( )3

6

RS O

SH SH

HN HN

HO

MPAA

O

HN

O

7

8

Figure 6 | Photodeprotection and sumoylation of folded histone H2B. (a) Scheme depicting photolytic cleavage of the auxiliary protecting group from H2B(A117C)photoaux (5) to generate H2B(A117C)aux (6). (b) ESI-MS

spectrum of 5. Calculated for 5, 14,059.2 Da. Observed for 5, 14,062.72.8 Da. (c) Circular dichroism spectrum of 5 in 50 mM Na2HPO4, pH 7.5. (d) ESI-MS spectrum of 6. Calculated for 6, 13,924.1 Da.

Observed for 6, 13,925.82.6 Da. (e) Circular dichroism spectrum of 6 in 50 mM Na2HPO4, pH 7.5. (f) Scheme depicting ligation of 6 to SUMO-3 (291)C47S-a-thioester to yield H2B(A117C)Su(C47S)aux (7) and subsequent MPAA-mediated auxiliary removal to yield H2B(A117C)Su(C47S)

(8) under folded conditions. (g) Coomassie-stained 15% SDS polyacrylamide gel electrophoresis gel of ligation between H2B(A117C)aux

(6) and SUMO3(291)C47S-a-thioester under non-denaturing conditions. Lane 1 SUMO-3(291)C47S-MES, 2 H2B(A117C)aux, 3 24 h ligation,

4 48 h ligation, 5 24 h MPAA incubation. (h) ESI-MS spectrum of the

ligation product, H2B(A117C)Su(C47S)aux (7). Calculated for 7, 24,225.6 Da.

Observed for 7, 24,230.94.0 Da. (i) ESI-MS spectrum of the nal product, H2B(A117C)Su(C47S) (8). Calculated for 8, 24,149.6 Da. Observed

for 8, 24,153.13.2 Da.

g

1 2 3 4 5

25 kDa

20 kDa

15 kDa

H2B(A117C)Su(C47S)

H2B(A117C)aux

SUMO-3(291)C47S

h i

7

8

+20

+23 +18

+16

+20

+23 +18

Intensity

+26

+26

+16

800 1,600 Mass (m/z)

800 1,600 Mass (m/z)

8 NATURE COMMUNICATIONS | 7:12979 | DOI: 10.1038/ncomms12979 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12979 ARTICLE

nal product, H2B(A117C)Su(C47S), was analysed by LC-ESI-MS, size exclusion chromatography and circular dichroism.

Thiyl radical detection by oxidation of NADH. To investigate the formation of thiyl radicals under auxiliary removal conditions, reduced NADH was dissolved at a nal concentration of 40 mM in solutions containing 200 mM of each thiol and 100 mM Na2HPO4, pH 7.3. The solutions were incubated at 25 C, protected from light. Absorbance at 340 nm was measured at various time points. A decrease in absorbance, due to oxidation of NADH, suggested the presence of thiyl radicals18. To further conrm the ability of aromatic thiols to perform single-electron transfer reactions, solutions were prepared containing 200 mM of each thiol and 100 mM Na2HPO4, pH 7.3, and these solutions were added to dry aliquots of the radical indicator, methyl viologen (MV2 ), for a nal concentration of 20 mM MV2 . Both

MV2 and its two-electron reduction product (MV) have absorbance maxima less than 400 nm. The single-electron reduction product MV , however, has a strong characteristic absorbance at 605610 nm (ref. 49). The resulting deep purple colour was observed immediately upon mixing methyl viologen with the aromatic thiol solutions, but no colour change occurred with aliphatic thiols, even after 24 h.

EPR experiments. A 200 mM MPAA stock solution was prepared in DMF. Aliquots from the stock solution were diluted with DMF and 50 mM Na2HPO4 at pH 7.5 to a nal 1:1 water-DMF mixture. Then, 100 mM DMPO was added, the samples were vortexed, transferred to a at cell and EPR spectra were recorded at room temperature on a Bruker EMX spectrometer equipped with a high sensitivity cavity and operating at 9.65 GHz and 100 KHz eld modulation. MPAA alkylation was performed by incubating 50 mM MPAA with 70 mM 2-iodoacetamide in50 mM Na2HPO4 at pH 7.5, in 1:1 water-DMF for 1.5 h at 25 C before DMPO addition. Parallel controls were also performed by pre-incubating 50 mM MPAA alone in 50 mM Na2HPO4 at pH 7.5, in a 1:1 water-DMF for 1.5 h at 25 C before adding DMPO. Computer simulation was performed using the Winsim program from P.E.S.T.23.

SENP1 hydrolysis assay. Size-exclusion-puried H2B(A117C)Su(C47S) (8) was assayed with the catalytic domain of sentrin-specic protease 1 (SENP1, Boston Biochem). SENP1 (0.05 nmol) was pre-activated in 10 ml buffer containing 50 mM tris, 150 mM NaCl, 12 mM dithiothreitol, pH 8 for 20 min at 25 C. To the reduced SENP1 was then added 10 ml of a solution containing 0.5 nmol of 8 in 50 mM tris, 150 mM NaCl, 1 mM dithiothreitol, pH 7.5. The resulting mixture was incubated for 24 h at 37 C. The assay was quenched by the addition of 6 Laemmli buffer

containing 300 mM dithiothreitol and boiled for 5 min, then run on an 18% SDSpolyacrylamide gel electrophoresis gel at 200 V for 1.5 h and stained with Coomassie brilliant blue.

Data availability. The data that support the ndings of this study are available with the article and its Supplementary Information Files, and from the corresponding author upon reasonable request.

References

1. Sharp, P. M. & Li, W. H. Molecular evolution of ubiquitin genes. Trends Ecol. Evol. 2, 328332 (1987).

2. Van der Veen, A. G. & Ploegh, H. L. Ubiquitin-like proteins. Annu. Rev. Biochem. 81, 323357 (2012).

3. Li, W. et al. Genome-wide and functional annotation of human E3 ubiquitin ligases identies MULAN, a mitochondrial E3 that regulates the organelles dynamics and signaling. PLoS ONE 3, e1487 (2008).

4. Komander, D. & Rape, M. The ubiquitin code. Annu. Rev. Biochem. 81, 203229 (2012).

5. Danielsen, J. M. R. et al. Mass spectrometric analysis of lysine ubiquitylation reveals promiscuity at site level. Mol. Cell Proteomics 10, M110.003590 (2011).

6. Weller, C. E., Pilkerton, M. E. & Chatterjee, C. Chemical strategies to understand the language of ubiquitin signaling. Biopolymers 101, 144155 (2014).

7. Abeywardana, T. & Pratt, M. R. Using chemistry to investigate the molecular consequences of protein ubiquitylation. ChemBioChem 15, 15471554 (2014).

8. Chatterjee, C., McGinty, R. K., Pellois, J. P. & Muir, T. W. Auxiliary-mediated site-specic peptide ubiquitylation. Angew. Chem. Int. Ed. Engl. 46, 28142818 (2007).

9. Wan, Q. & Danishefsky, S. J. Free-radical-based, specic desulfurization of cysteine: a powerful advance in the synthesis of polypeptides and glycopolypeptides. Angew. Chem. Int. Ed. Engl. 46, 92489252 (2007).10. Kumar, K. S. A., Harpaz, Z., Haj-Yahya, M. & Brik, A. Side-chain assisted ligation in protein synthesis. Bioorg. Med. Chem. Lett. 19, 38703874 (2009).

11. Merkx, R. et al. Scalable synthesis of g-thiolysine starting from lysine and a side by side comparison with d-thiolysine in non-enzymatic ubiquitination. Chem. Sci. 4, 44944498 (2013).

12. Canne, L. E., Bark, S. J. & Kent, S. B. H. Extending the applicability of native chemical ligation. J. Am. Chem. Soc. 118, 58915896 (1996).

13. Weller, C. E., Huang, W. & Chatterjee, C. Facile synthesis of nativeand protease-resistant ubiquitylated peptides. ChemBioChem 15, 12631267 (2014).

14. Johnson, E. C. B. & Kent, S. B. H. Insights into the mechanism and catalysis of the native chemical ligation reaction. J. Am. Chem. Soc. 128, 66406646 (2006).

15. Wang, C., Guo, Q. X. & Fu, Y. Theoretical analysis of the detailed mechanism of native chemical ligation reactions. Chem. Asian J. 6, 12411251 (2011).

16. Jin, W. et al. A unique class of duocarmycin and CC-1065 analogues subject to reductive activation. J. Am. Chem. Soc. 129, 1539115397 (2007).

17. Fava, A., Reichenbach, G. & Peron, U. Kinetics of the thiol-disulde exchange.II. Oxygen-promoted free-radical exchange between aromatic thiols and disuldes. J. Am. Chem. Soc. 89, 66966700 (1967).18. Forni, L. G. & Willson, R. L. Thiyl and phenoxyl free radicals and NADH: direct observation of one-electron oxidation. Biochem. J. 240, 897903 (1986).

19. Zhao, R., Lind, J., Merenyi, G. & Eriksen, T. E. Kinetics of one-electron oxidation of thiols and hydrogen abstraction by thiyl radicals from a-amino

C-H bonds. J. Am. Chem. Soc. 116, 1201012015 (1994).20. Jeschke, G. EPR techniques for studying radical enzymes. Biochim. Biophys. Acta 1707, 91102 2005.

21. Augusto, O., Bonini, M. G. & Trindade, D. Spin trapping of glutathiyl and protein radicals produced from nitric oxide-derived oxidants. Free Radic. Biol. Med. 36, 12241232 (2004).

22. Pou, S. & Rosen, G. M. Generation of thiyl radical by nitric oxide: a spin trapping study. J. Chem. Soc., Perkin Trans. 2, 15071512 (1998).

23. Duling, D. R. Simulation of multiple isotropic spin-trap EPR spectra. J. Magn. Reson. B 104, 105110 (1994).

24. Mile, B., Rowlands, C. C., Sillman, P. D. & Fildes, M. The EPR spectra of thiyl radical spin adducts produced by photolysis of disuldes in the presenceof 2,4,6-tri-tert-butynitrosobenrene and 5,5-dimethyl-1-pyrroline N-oxide.J. Chem. Soc. Perkin Trans. 2, 14311437 (1992).25. Wu, M. & Begley, T. P. b-scission of the N O bond in alkyl hydroxamate

radicals: a fast radical trap. Org. Lett. 2, 13451348 (2000).26. Gaspari, G. & Granzow, A. The ash photolysis of mercaptans in aqueous solution. J. Phys. Chem. 74, 836839 (1970).

27. Bitter, F. High-eld EPR detection of a disulde radical anion in the reduction of cytidine 5-diphosphate by the E441Q R1 mutant of Escherichia coli ribonucleotide reductase. Proc. Natl Acad. Sci. USA 96, 89798984 (1999).

28. Tung, T.-L. & Stone, J. A. The formation and reactions of disulde radical anions in aqueous solution. Can. J. Chem. 53, 31533157 (1975).

29. Thompson, R. E. et al. Triuoroethanethiol: an additive for efcient one-pot peptide ligation-desulfurization chemistry. J. Am. Chem. Soc. 136, 81618164 (2014).

30. Petlicki, J. & van de Ven, T. G. M. The equilibrium between the oxidation of hydrogen peroxide by oxygen and the dismutation of peroxyl or superoxide radicals in aqueous solutions in contact with oxygen. J. Chem. Soc., Faraday Trans. 94, 27632767 (1998).

31. Kuppusamy, P. & Zweier, J. L. Characterization of free radical generation by xanthine oxidase. J. Biol. Chem. 264, 98809884 (1989).

32. Purdie, J. W., Gillis, H. A. & Klassen, N. V. Pulse radiolysis of penicillamine in aqueous solution: the thiyl radical and the disulphide radical anion. Chem. Commun. 51, 11631165 (1971).

33. Antonello, S., Daasbjerg, K., Jensen, H., Taddei, F. & Maran, F. Formation and cleavage of aromatic disulde radical anions. J. Am. Chem. Soc. 125, 1490514916 (2003).

34. Goldstein, S., Meyerstein, D. & Czapski, G. The Fenton reagents. Free Radic. Biol. Med. 15, 435445 (1993).

35. Paulsen, C. E. & Carroll, K. S. Cysteine-mediated redox signaling: chemistry, biology, and tools for discovery. Chem. Rev. 113, 46334679 (2013).

36. Gaussian Development Version. Revision H.36 v. 9.0 (Gaussian Inc., 2014).37. Moyal, T., Hemantha, H. P., Siman, P., Refua, M. & Brik, A. Highly efcient one-pot ligation and desulfurization. Chem. Sci. 4, 24962501 (2013).

38. Shiio, Y. & Eisenman, R. N. Histone sumoylation is associated with transcriptional repression. Proc. Natl Acad. Sci. USA 100, 1322513230 (2003).

39. Dhall, A. et al. Sumoylated human histone H4 prevents chromatin compaction by inhibiting long-range internucleosomal interactions. J. Biol. Chem. 289, 3382733837 (2014).

40. Zheng, J. S., Tang, S., Qi, Y. K., Wang, Z. P. & Liu, L. Chemical synthesis of proteins using peptide hydrazides as thioester surrogates. Nat. Protoc. 8, 24832495 (2013).

41. Lang, K. & Chin, J. W. Cellular incorporation of unnatural amino acids and bioorthogonal labeling of proteins. Chem. Rev. 114, 47644806 (2014).

42. Hendriks, I. A. et al. Uncovering global SUMOylation signaling networks in a site-specic manner. Nat. Struct. Mol. Biol. 21, 927936 (2014).

43. Chandrasekharan, M. B., Huang, F. & Sun, Z. W. Ubiquitination of histone H2B regulates chromatin dynamics by enhancing nucleosome stability. Proc. Natl Acad. Sci. USA 106, 1668616691 (2009).

NATURE COMMUNICATIONS | 7:12979 | DOI: 10.1038/ncomms12979 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12979

44. Mattingly, P. G. & Miller, M. J. Titanium trichloride reduction of substituted N-hydroxy-2-azetidinones and other hydroxamic acids. J. Org. Chem. 45, 410415 (1980).

45. Denmark, S. E., Stolle, A. & Dixon, J. A. Tandem inter [4 2]/Intra

[3 2] cycloadditions. 6. The bridged mode. J. Am. Chem. Soc. 117, 21002101

(1995).46. Keck, G. E., Fleming, S., Nickell, D. & Weider, P. Mild and efcient methods for the reductive cleavage of nitrogen-oxygen bonds. Syn. Commun. 9, 281286 (1979).

47. Keck, G. E., Wager, T. T. & McHardy, S. F. Reductive cleavage of N-o bonds in hydroxylamines and hydroxamic acid derivatives using samarium diiodide. Tet. Lett. 55, 1175511772 (1999).

48. Cutulic, S., Murphy, J., Farwaha, H., Zhou, S.-Z. & Chrystal, E. Metal-Free reductive cleavage of N-O bonds in weinreb amides by an organic neutral super-electron donor. Synlett 2008, 21322136 (2008).

49. Watanabe, T. & Honda, K. Measurement of the extinction coefcient of the methyl viologen cation radical and the efciency of its formation by semiconductor photocatalysis. J. Phys. Chem. 86, 26172619 (1982).

Acknowledgements

We would like to thank Professors Derek Pratt, Forrest Michael and Paul Hopkins for valuable discussions. C.E.W. gratefully acknowledges support from an NSF GRFP (grant number DGH-1256082) and an ARCS foundation fellowship. C.C. acknowledges support from the NIH/NIGMS (grant 1R01GM110430) and the Department of Chemistry at the University of Washington, Seattle. O.A. acknowledges support from the FAPESP (grant number 2013/07937-8). X.L. acknowledges support from the NSF (grant number CHE- 1565520).

Author contributions

C.E.W. performed all protein ligation and reduction reactions. C.E.W., A.D., S.D.W., N.A.S., E.L.T. and J.D.B. contributed new reagents. F.D. performed all computational calculations. E.L. and O.A. performed all EPR experiments. C.E.W., F.D., and C.C. wrote the paper. All authors read the paper and provided comments.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Weller, C. E. et al. Aromatic thiol-mediated cleavage of NO bonds enables chemical ubiquitylation of folded proteins. Nat. Commun. 7,12979doi: 10.1038/ncomms12979 (2016).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

r The Author(s) 2016

10 NATURE COMMUNICATIONS | 7:12979 | DOI: 10.1038/ncomms12979 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications